JP2006180971A - Skin evaluation method and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a skin evaluation method and device capable of evaluating the state of the skin even if a sampled image is ambiguous on some level and individual differences of image sampling devices are large. <P>SOLUTION: A part of an image input from a fingerprint sensor (S1) is picked up (S3) and a fundamental frequency is calculated using a secondary linear prediction analysis (S5). The sum S of an x-directional fundamental frequency and y-directional fundamental frequency is calculated and defined as a determination criterion of the fineness of the texture (S7). A prepared threshold and the criterion S found in S7 are compared with each other and the fineness of the texture of the skin state is evaluated by being classified into three levels (S9-S17). A ratio of the x-directional to y-directional fundamental frequencies of an object image is calculated and defined as an evaluation criterion on the streak of the texture (S19), the image is rotated to find a similar evaluation criterion, and the streak of the texture is determined based thereon by being classified into two levels (S31-S35). The determination result of the fineness and the streak of the texture are displayed on the display screen (S37). <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a skin evaluation method and a skin evaluation apparatus.

  In general, it is a desire of many people to obtain healthy and beautiful skin. In order to fulfill the desire to be beautiful, many companies have developed and sold a wide variety of cosmetics and beauty equipment. Cosmetics and beauty instruments are compatible with each other, and in order for these to exhibit a sufficient effect, it is necessary to select a beauty method suitable for different skin qualities and skin conditions at that time. Therefore, the diagnosis of skin condition is very important.

  Here, the state of the skin (hereinafter sometimes referred to as “skin”) is generally evaluated by “texture”. The texture is a mesh-like unevenness spreading on the surface of the skin, and is formed by the “skin”, which is a high part, and the “skin” and “pore” between the hides. “Skin texture” is generally evaluated by “texture fineness” and “texture shape”. “Smooth skin” is a state in which beautiful and uniform undulations continue. In addition, a state where the skin is streaked is a state where “texture is flowing”. The textured shape is good when the texture is well shaped and the lines are not visible.

  Until a while ago, skin conditions were diagnosed only by specialists in cosmetic companies. However, the expert's intuitive judgment is an important factor for skin diagnosis by specialists, so evaluation requires long-term proficiency and is not easily performed by anyone. In recent years, in order to solve this problem, a skin diagnosis apparatus that automatically performs skin diagnosis has been devised.

  For example, in Patent Document 1, (1) an RGB image (color image) of skin is obtained. (2) Conversion from RGB to Lab image. (3) A variation coefficient (variation) of the area or position of the entire image of the low luminance region is calculated. (4) The unevenness of the skin is evaluated using this coefficient of variation. In this way, we have proposed a skin diagnostic method that uses skin non-uniformity as an evaluation criterion.

In Patent Document 2, the point where three or more skin grooves meet on the skin surface is defined as one intersection, and the roughness of the skin texture is measured by measuring the number of intersections per unit area and the average area per intersection. A method of measuring is disclosed.
Japanese Patent No. 3351958 JP 2001-170028 A

  However, since the above-mentioned Patent Document 1 is based on the premise that the skin image to be analyzed has a relatively good image quality, it is necessary to use a relatively expensive camera capable of obtaining a color image. Further, a threshold value for determining the low luminance area is required. There is a problem that it is difficult to set a threshold unless a relatively expensive camera with little individual difference is used.

  In Patent Document 2, binarization and thinning are required. The use of low-individual cameras that can set a binarization threshold, or skin image collection under certain conditions is assumed, and it is difficult to analyze low-quality images obtained with inexpensive cameras. There is.

  The present invention has been made to solve the above problem, and a skin evaluation method capable of evaluating the state of the skin even when the collected image is somewhat unclear or when there is a large individual difference between the image collection devices. And it aims at providing a skin evaluation apparatus.

  In order to achieve the above object, a skin evaluation method according to claim 1 of the present invention performs frequency analysis on an input skin image, and based on frequency characteristics of the skin image obtained as a result of the frequency analysis. The skin condition is determined.

  Moreover, the skin evaluation method according to claim 2 of the present invention extracts the fundamental frequency of the skin image as the frequency feature by the frequency analysis in addition to the configuration of the invention according to claim 1, and the fundamental frequency Determines that the condition of the skin is good. Here, the fundamental frequency referred to in the present invention includes a fundamental frequency obtained by a well-known frequency analysis, as well as a fundamental frequency obtained by a zero-cross method, which is a simple fundamental frequency calculation method.

  In addition to the configuration of the invention described in claim 1 or 2, the skin evaluation method according to claim 3 of the present invention performs the frequency analysis for each of the X direction and the Y direction of the skin image. The fundamental frequency is extracted for each of the X direction and the Y direction of the skin image, a ratio of the fundamental frequency in the X direction and the fundamental frequency in the Y direction is calculated, and the ratio is within a predetermined threshold range. It is determined that the skin condition is good.

  A skin evaluation method according to claim 4 of the present invention uses a second-order linear prediction analysis as the frequency analysis in addition to the configuration of the invention according to any one of claims 1 to 3.

  According to a skin evaluation method of a fifth aspect of the present invention, in addition to the configuration of the invention according to any one of the first to fourth aspects, the skin image is input by a fingerprint sensor.

  Furthermore, the skin evaluation apparatus according to claim 6 of the present invention is an image input means for inputting a skin image, a frequency analysis means for frequency analysis of the skin image input by the image input means, and the frequency analysis means. And a feature extracting unit that extracts a frequency feature of the skin image obtained as a result of the frequency analysis by the method and a determining unit that determines the state of the skin based on the frequency feature extracted by the feature extracting unit. .

  Further, in the skin evaluation apparatus according to claim 7 of the present invention, in addition to the configuration of the invention according to claim 6, the feature extraction means extracts the fundamental frequency of the skin image as the frequency feature, and the determination The means determines that the skin condition is good when the fundamental frequency exceeds a predetermined threshold value.

  Further, in the skin evaluation apparatus according to claim 8 of the present invention, in addition to the configuration of the invention according to claim 6 or 7, the frequency analysis means performs frequency analysis on each of the X direction and the Y direction of the skin image. The feature extraction unit extracts the fundamental frequency for each of the X direction and the Y direction of the skin image, and the ratio of the fundamental frequency in the X direction and the fundamental frequency in the Y direction extracted by the feature extraction unit. Frequency ratio calculating means for calculating the ratio, and the determining means determines that the skin condition is good if the ratio calculated by the frequency ratio calculating means is within a predetermined threshold range.

  According to a ninth aspect of the present invention, in addition to the configuration of the sixth aspect, the frequency analysis means uses a second-order linear prediction analysis.

  According to a tenth aspect of the present invention, in addition to the configuration of the sixth aspect of the present invention, the image input means includes a fingerprint sensor.

  The skin evaluation method according to claim 1 of the present invention performs frequency analysis on the input skin image, and determines the skin state based on the frequency characteristics of the skin image obtained as a result of the frequency analysis. Such a frequency feature can be obtained even if the input skin image is not clear, so that the skin condition can be evaluated regardless of the image quality of the input skin image.

  The skin evaluation method according to claim 2 of the present invention extracts the fundamental frequency as a frequency feature in addition to the effect of the invention according to claim 1. The correlation that the skin state is good when the fundamental frequency is high, that is, the so-called fine state, is confirmed, and the skin state can be quantified and determined by the fundamental frequency.

  Moreover, the skin evaluation method according to claim 3 of the present invention, in addition to the effect of the invention according to claim 1 or 2, determines the flow of skin texture when the ratio of the fundamental frequency in the X direction and the Y direction is obtained. The skin condition can be evaluated from the texture shape according to the degree of the flow. In particular, by evaluating together with the fineness of the texture obtained in claim 2, higher quality skin evaluation becomes possible.

  The skin evaluation method according to claim 4 of the present invention uses a second-order linear prediction analysis as frequency analysis in addition to the effect of the invention according to any one of claims 1 to 3, so that voice or fingerprint authentication is used. Skin can be evaluated by techniques well known in the field.

  In addition to the effect of the invention according to any one of claims 1 to 4, the skin evaluation method according to claim 5 of the present invention inputs a skin image using a fingerprint sensor. Compared with the case where it does, the mechanism which makes the imaging distance of skin and a camera lens constant, the effort to focus, etc. are unnecessary, and the skin image made into evaluation object can be obtained easily.

  Furthermore, the skin evaluation apparatus according to claim 6 of the present invention performs frequency analysis on the input skin image, and determines the skin state based on the frequency characteristics of the skin image obtained as a result of the frequency analysis. Such a frequency feature can be obtained even if the input skin image is not clear, so that the skin condition can be evaluated regardless of the image quality of the input skin image.

  Moreover, the skin evaluation apparatus according to claim 7 of the present invention extracts the fundamental frequency as a frequency feature in addition to the effect of the invention according to claim 6. The correlation that the skin state is good when the fundamental frequency is high, that is, the so-called fine state, is confirmed, and the skin state can be quantified and determined by the fundamental frequency.

  In addition to the effect of the invention according to claim 6 or 7, the skin evaluation apparatus according to claim 8 of the present invention obtains the flow of skin texture when the ratio of the fundamental frequencies in the X direction and the Y direction is obtained. The skin condition can be evaluated from the texture shape according to the degree of the flow. In particular, by evaluating together with the fineness of the texture obtained in claim 7, higher quality skin evaluation becomes possible.

  Moreover, since the skin evaluation apparatus according to claim 9 of the present invention uses a second-order linear prediction analysis as frequency analysis in addition to the effects of the invention according to any of claims 6 to 8, voice and fingerprint authentication Skin can be evaluated by techniques well known in the field.

  In addition to the effect of the invention according to any one of claims 6 to 9, the skin evaluation apparatus according to claim 10 of the present invention inputs a skin image using a fingerprint sensor. Compared with the case where it does, the mechanism which makes the imaging distance of skin and a camera lens constant, the effort to focus, etc. are unnecessary, and the skin image made into evaluation object can be obtained easily.

  Next, embodiments to which the present invention is applied will be described below with reference to the drawings. In the following embodiments, a program for causing a computer to execute the skin evaluation method of the present invention is installed in a mobile phone with a fingerprint sensor, and this mobile phone functions as the skin evaluation apparatus of the present invention. The outline of this embodiment will be described. When a skin evaluation program installed in a mobile phone is started and an operator reads a skin image by a fingerprint sensor mounted on the mobile phone, the skin of the input skin image is read by the skin evaluation program. Is evaluated and the result is displayed on the display screen.

  First, the configuration of the mobile phone will be described with reference to FIGS. FIG. 1 is an external view of the mobile phone 1. FIG. 2 is a block diagram showing an electrical configuration of the mobile phone 1.

  As shown in FIG. 1, the mobile phone 1 includes a display screen 2, a numeric key input unit 3, a jog pointer 4, a call start button 5, a call end button 6, a microphone 7, and a speaker 8. , Function selection buttons 9 and 10, a fingerprint sensor 11 as an image input means, and an antenna 12 (see FIG. 2) are provided. The numeric key input unit 3, the jog pointer 4, the call start button 5, the call end button 6, and the function selection buttons 9 and 10 constitute a key input unit 38 (see FIG. 2).

  Here, the fingerprint sensor 11 may be a capacitance type sensor, an optical sensor, a heat sensitive type, an electric field type, a planar type, or a line type fingerprint sensor. It suffices if the whole can be acquired as fingerprint information. In this embodiment, it is assumed that a line-type fingerprint sensor is used, and the skin image is read from the fingerprint sensor 11 by the operator holding the mobile phone 1 and sliding the fingerprint sensor 11 along the skin.

  As shown in FIG. 2, the cellular phone 1 includes an analog front end 36 that performs amplification of an audio signal from the microphone 7 and amplification of audio output from the speaker 8, and audio amplified by the analog front end 36. An audio codec unit 35 that converts the signal into a digital signal and converts the digital signal received from the modem 34 into an analog signal so that it can be amplified by the analog front end 36, a modem unit 34 that performs modulation / demodulation, and amplification of radio waves received from the antenna 12 A transmission / reception unit 33 that performs detection and modulates and amplifies the carrier signal with the signal received from the modem 34 is provided.

  Further, the mobile phone 1 is provided with a control unit 20 that controls the entire mobile phone 1. The control unit 20 includes a CPU 21, a RAM 22 that temporarily stores data, and a clock function unit 23. ing. The RAM 22 is used as a work area in processing to be described later, and stores an area for storing a value obtained by converting the unevenness of the skin acquired from the fingerprint sensor 11 into a density value, and a calculation result calculated in each processing to be described later. A storage area such as a storage area is prepared. The control unit 20 is connected to a key input unit 38 for inputting characters and the like, the display screen 2, the fingerprint sensor 11, the nonvolatile memory 30, and a melody generator 32 for generating a ring tone. The melody generator 32 is connected to a speaker 37 that utters a ring tone generated by the melody generator 32. The nonvolatile memory 30 is provided with an area for storing various programs executed by the CPU 21 of the control unit 20, an area for storing various initial setting values, an area for storing various predetermined threshold values, and the like.

  Next, the skin evaluation process executed in the mobile phone 1 having the above configuration will be described with reference to FIGS. FIG. 3 is a flowchart showing the flow of the skin evaluation process. FIG. 4 is an explanatory diagram showing a skin image input by the fingerprint sensor and a sample of a small area cut out from the skin image. FIG. 5 is a graph showing the fundamental frequency extracted from the skin image. FIG. 6 is an explanatory diagram showing a sample of the skin image and the fundamental frequency. FIG. 7 is an explanatory diagram showing a sample of a skin image and a fundamental frequency when obtaining a texture flow evaluation scale. FIG. 8 is an explanatory diagram showing a sample display screen showing the skin evaluation determination result.

  As shown in FIG. 3, when the skin evaluation process is started, first, a skin image 100 as shown in FIG. 4 input from the fingerprint sensor 11 is acquired (S1). As shown in FIG. 4, the size of the skin image acquired in this embodiment is 224 pixels wide and the height is arbitrary (H pixels). Next, from the acquired skin image, a small region 101 having the darkest gradation value is cut out to be an evaluation target (S3). The size of the small region cut out here is 128 pixels × 128 pixels. However, the present invention is not limited to this. Next, the fundamental frequency (skin pitch) of the cut out small area is obtained (S5). Here, the fundamental frequency is obtained by the following method using second-order linear prediction analysis.

First, the X-direction waveform 1 line Fi, j (i = 0, 1,... M−1, j = 0, 1,... N−1 of the small region image cut out in S3: where M is Hi, j is obtained by multiplying the number of pixels in the y direction (128 in this embodiment) and N is the number of pixels in the x direction (128 in this embodiment) by applying a Hamming window as shown in the following equation (1).

Next, autocorrelation coefficients ri, 1 and ri, 2 are obtained from the obtained Hi, j as shown in the following equation (2).

Then, linear prediction coefficients αi, 0 and αi, 1 are obtained from the autocorrelation coefficients ri, 1 and ri, 2 by the following equation (3).

Finally, a normalized resonance frequency fi normalized from 0 to π is calculated from the linear prediction coefficients αi, 0 and αi, 1 by the following equation (4), and is set as a fundamental frequency in the x direction.

  Since the fundamental frequency fi in the x direction is calculated by the above equations (2) to (4), the linear prediction coefficients αj, 0 and αj, 1 are similarly calculated from the autocorrelation coefficients rj, 1 and rj, 2. And the normalized resonance frequency fj is calculated to obtain the fundamental frequency in the y direction. The fundamental frequency thus obtained is as shown in FIG.

  Here, the fundamental frequency is obtained by the second-order linear prediction analysis, but the fundamental frequency may be calculated using the autocorrelation coefficient obtained by the above equation (2) as it is. Further, the fundamental frequency may be calculated using a zero cross method widely used for simple frequency analysis. In the case of using the zero cross method, it is preferable to obtain a zero cross number that is the number of times that the number of pixels of one line of the skin image to be analyzed intersects a threshold value or a fixed value obtained in advance by experiment, and this is used as a fundamental frequency. If the number of zero crossings is large, the fundamental frequency is high.

There is a correlation that the fundamental frequency obtained as described above becomes higher when the texture of the skin image is fine and becomes lower when the texture is coarse. Therefore, if the total of the 128 lines of fundamental frequencies calculated in this embodiment is obtained, the texture is fine if the total value is high, and the texture is coarse if the total value is low. It can be used as a judgment scale. Therefore, next, a total value S of the fundamental frequency fx (i) in the x direction and the fundamental frequency fy (i) in the y direction is calculated by the following equation (5) (S7).

  In the above formula (5), the determination scale is the sum of the fundamental frequency in the x direction and the fundamental frequency in the y direction. However, the determination scale may be calculated using only the fundamental frequency in the x direction and the fundamental frequency in the y direction. .

  Next, the fineness of the texture is evaluated using the texture determination scale obtained in S7. In this embodiment, the fineness of skin texture is evaluated in three stages, and two threshold values are prepared. For example, threshold 1 is set to 60 and threshold 2 is set to 40. If the texture determination scale is higher than the threshold value 1 or higher, the texture is fine and the skin condition is good. If it is between the threshold value 1 and the threshold value 2, it is medium, and if it is below the threshold value 2, the texture is not much. It can be evaluated that it is not fine and requires skin care.

  Therefore, first, it is determined whether or not the threshold value is 1 or more (S9). If the texture determination scale is greater than or equal to the threshold value 1 (S9: YES), the evaluation regarding the texture fineness is temporarily stored in the RAM 22 as “good” (S11). The relationship between the skin image evaluated as “good” and the texture determination scale S is, for example, data01, 02, 03, 04 in FIG. In any case, the texture determination scale S exceeds 60, and it is understood that the texture is well-organized by visual observation of the state of the skin image.

  If the texture determination scale is not greater than or equal to the threshold 1 (S9: NO), it is next determined whether or not the texture determination scale is greater than or equal to the threshold 2 (S13). If the texture determination scale is greater than or equal to the threshold 2 (S13: YES), the evaluation regarding the texture fineness is temporarily stored in the RAM 22 as "medium" (S15). The relationship between the skin image evaluated as “medium” and the texture determination scale S is, for example, data 05, 06, 07 in FIG. In any case, the texture determination scale S is between 40 and 60. When the state of the skin image is visually observed, it can be said that the texture is not perfect, but it cannot be said that it is rough.

  If the texture determination scale is not greater than or equal to the threshold value 2 (S13: NO), the evaluation regarding the texture fineness is temporarily stored in the RAM 22 as “bad” (S17). The relationship between the skin image evaluated as “bad” and the texture determination scale S is, for example, data08, 09, 10, 11, 12 in FIG. In any case, the texture determination scale S is less than 40, and it can be seen from the visual observation of the state of the skin image that the texture is rough. As described above, there is a correlation between the skin condition and the determination scale S.

Next, the texture shape is evaluated. The texture shape evaluation is performed by calculating the ratio of the fundamental frequency in the x and y directions of the target image obtained in S3. When the fundamental frequency in the y direction is larger than the fundamental frequency in the x direction, the texture flows sideways. For example, in the skin image 110 and the skin image 111 in FIG. 7, the value of the fundamental frequency Sy in the y direction is larger than the value of the fundamental frequency Sx in the x direction, and it is observed that the texture flows laterally from the image. it can. On the other hand, when the fundamental frequency Sy in the y direction is smaller than the fundamental frequency Sx in the x direction, the texture tends to flow vertically. Using this property, an evaluation scale for determining the texture flow is obtained by the following equation (6) (S19).
Na = MAX (Sx, Sy) / MIN (Sx, Sy) (6)
The evaluation scale Na calculated by the above equation (6) is the ratio of the fundamental frequency in the x direction and the y direction. When Na≈1, the texture is in place. It can be determined that texture is flowing when Na >> 1.

  However, in this method, when the flow direction has an inclination close to 45 °, Sx and Sy have similar values, so that it is difficult to detect the flow. In order to avoid this problem, the pitch ratio may be obtained again from the image obtained by rotating the cutout image. The rotated image may be only a 45-degree rotated image, or an image rotated by 22.5 °, 45 °, or 67.5 ° may be used, for example, in order to increase accuracy.

  Therefore, after calculating the original image flow evaluation scale in S19, it is determined whether or not the image rotation processing for flow determination is completed (S21). Then, S23, S25, and S19 are repeated according to the number of images to be rotated. If the image rotation processing for flow determination has not been completed yet (S21: NO), the target skin image is rotated by a predetermined angle as shown in FIG. 7 to obtain a rotated image 112 (S23). And the fundamental frequency about x direction and y direction is calculated about the rotation image 112 (S25). A state in which the fundamental frequency is obtained for the rotated image is an image 113 shown in FIG.

Next, returning to S19, the evaluation scale Na for the rotated image at the angle is obtained by the above equation (6). Then, it is determined whether or not the processing for all the scheduled rotation images has been completed (S21). When the evaluation scale Na is obtained for all the rotated images (S21: YES), the evaluation scale N0 for the original image and the evaluation scale for each round image, for example, 22.5 °, 45 °, 67.5 ° For example, based on N22.5, N45, and N67.5, the evaluation scale is calculated again by the following formula (7) (S29).
N = MAX (N0, N22.5, N45, N67.5) (7)

  In the above formula (7), the maximum value of the evaluation scale Na obtained for each image is adopted as the evaluation scale N for the flow of texture. This is calculated by obtaining the average value or the total value of each evaluation scale Na. May be. Further, when obtaining the maximum from N0, N22.5, N45, and N67.5, it is conceivable to improve the calculation accuracy of N by complementing these four points with a Gaussian window or the like. When mounted on a computer with a high processing speed, for example, the image can be rotated every 1 ° to improve the calculation accuracy of N.

  Next, the texture shape is evaluated using the texture flow evaluation scale obtained in S29. In the present embodiment, the evaluation of the texture of the skin is a two-stage evaluation of whether the texture is in place or flowing. Since the evaluation scale obtained in S29 is the ratio of the fundamental frequency in the x direction and the y direction, it is determined whether or not this evaluation scale is an approximate value of 1 (S31). When the evaluation scale is an approximate value of 1 (S31: YES), the texture shape evaluation is temporarily stored in the RAM 22 as “prepared” (S33). If the evaluation scale is not an approximate value of 1 (S31: NO), the texture shape evaluation is temporarily stored in the RAM 22 as "flowing" (S35).

  Through the above processing, the evaluation of the fineness of the texture and the evaluation of the texture shape are executed and stored in the RAM 22, so that both evaluation results are called up and displayed on the display screen 2 of the mobile phone 1 as shown in FIG. A result display screen is displayed (S37). FIG. 8 shows an example in which the fineness of the texture is medium and the texture is evaluated to be in good shape. As shown in this display screen, processing for executing comprehensive evaluation that combines evaluation of fineness of texture and shape of texture may be added. In this embodiment, the overall evaluation is displayed in three stages of A (good), B (medium), and C (defect = needs care).

  As described above, according to the mobile phone 1 that functions as the skin evaluation apparatus of the present embodiment, the basic frequency of the image read by the fingerprint sensor 11 is calculated to determine the fineness of the texture, and further, the skin image The shape of the texture is determined by calculating the ratio of the fundamental frequency in the x and y directions. Since the fundamental frequency can be calculated even when the skin image is not clear, it is possible to easily evaluate the skin condition without requiring an expensive device such as a camera. Further, since the processing load is relatively light, it is possible to perform processing at high speed even when it is incorporated in a device having the same CPU capability as a mobile phone, which is preferable.

  In the above embodiment, the CPU 21 that executes the fundamental frequency calculation process in S5 of the flowchart of FIG. 3 functions as the frequency analysis means of the present invention. In addition, the CPU 21 that executes the texture determination scale calculation process in S7 of the flowchart of FIG. 3 functions as the feature extraction unit of the present invention. Further, the CPU 21 that executes the texture fineness determination process in S9 and S13 in the flowchart of FIG. 3 and executes the texture shape determination process in S31 functions as the determination means of the present invention. Further, the CPU 21 that executes the flow evaluation scale calculation processing in S19 of the flowchart of FIG. 3 functions as the frequency ratio calculation means of the present invention.

  In the above embodiment, a program for causing a computer to execute the skin evaluation method of the present invention is incorporated in a mobile phone and functions as a skin evaluation apparatus. However, the embodiment of the present invention is not limited to the above configuration. Alternatively, it may be read and executed as an application program for a personal computer. Further, the input of the skin image is not limited to the fingerprint sensor, and may be captured by a camera.

1 is an external view of a mobile phone 1. FIG. 2 is a block diagram showing an electrical configuration of the mobile phone 1. FIG. It is a flowchart which shows the flow of a skin evaluation process. It is explanatory drawing which shows the sample of the small area cut out from the skin image input by the fingerprint sensor, and the skin image. It is a graph which shows the fundamental frequency extracted from the skin image. It is explanatory drawing which shows the sample of a skin image and a fundamental frequency. It is explanatory drawing which shows the sample of the skin image at the time of calculating | requiring the evaluation scale of a texture flow, and a fundamental frequency. It is explanatory drawing which shows the sample of the display screen showing the determination result of skin evaluation.

Explanation of symbols

1 Cellular Phone 2 Display Screen 11 Fingerprint Sensor 21 CPU
22 RAM
20 Control unit 30 Non-volatile memory 100 Skin image 101 Small region

Claims (10)

  A skin evaluation method for performing frequency analysis on an input skin image and determining the skin state based on frequency characteristics of the skin image obtained as a result of the frequency analysis.   The frequency analysis extracts the fundamental frequency of the skin image as the frequency feature, and determines that the skin condition is good when the fundamental frequency exceeds a predetermined threshold value. The skin evaluation method according to claim 1.   By performing the frequency analysis for each of the X direction and the Y direction of the skin image, the fundamental frequency is extracted for each of the X direction and the Y direction of the skin image, and the fundamental frequency in the X direction and the fundamental in the Y direction are extracted. The skin evaluation method according to claim 1 or 2, wherein a frequency ratio is calculated, and the skin condition is determined to be good if the ratio is within a predetermined threshold range.   The skin evaluation method according to claim 1, wherein a second-order linear prediction analysis is used as the frequency analysis.   The skin evaluation method according to claim 1, wherein the skin image is input by a fingerprint sensor. An image input means for inputting an image of the skin;
Frequency analysis means for frequency analysis of the skin image input by the image input means;
Feature extraction means for extracting frequency features of the skin image obtained as a result of frequency analysis by the frequency analysis means;
A skin evaluation apparatus comprising: a determination unit that determines the state of the skin based on the frequency feature extracted by the feature extraction unit. The feature extraction means extracts the fundamental frequency of the skin image as the frequency feature,
The skin evaluation apparatus according to claim 6, wherein the determination unit determines that the skin condition is good when the fundamental frequency exceeds a predetermined threshold value. The frequency analysis means performs frequency analysis for each of the X direction and the Y direction of the skin image,
The feature extraction means extracts the fundamental frequency for each of the X direction and the Y direction of the skin image,
Frequency ratio calculating means for calculating a ratio of the fundamental frequency in the X direction and the fundamental frequency in the Y direction extracted by the feature extracting means;
The skin according to claim 6 or 7, wherein the determination means determines that the skin condition is good if the ratio calculated by the frequency ratio calculation means is within a predetermined threshold range. Evaluation device.   The skin evaluation apparatus according to claim 6, wherein the frequency analysis unit uses a second-order linear prediction analysis. The skin evaluation apparatus according to claim 6, wherein the image input unit is a fingerprint sensor.